A Low Latency SCAN-Flip Polar Decoder for 5G Vehicular Communication

Abstract

Polar codes are widely considered as one of the most promising channel codes for future wireless communication. However, at short or moderate block lengths, their error-correction performance under traditional successive cancellation (SC) decoding is inferior to other modern channel codes, while under list decoding outperforms at the cost of high complexity and long latency. Successive cancellation flip (SCF) decoding is shown having competitive performance compared to that of list decoding but suffers from a long decoding latency. In this work, we propose the SCAN-Flip decoding algorithm by introducing the flipping idea into soft cancellation (SCAN) decoding. The proposed algorithm improves the error-correction performance of soft cancellation decoding and accelerates the convergence of iterative calculation, leading to lower execution-time. Besides, we also propose a new path metric to improve the performance of our SCAN-Flip decoder further. Simulation results show that the proposed decoder has a much smaller average number of iterations than that of SCF at equivalent frame error rate. At equivalent max number of iterations, the error-correction performance of SCAN-Flip outperforms SC-Flip by up to 0.25 dB at bit error rate of $10^{-4}$.

Supported by the NSF of China (Grant No. 61170083) and Specialized Research Fund for the Doctoral Program of Higher Education (Grant No. 20114307110001).